LAPS_Spring2018_LAPS_Fall13_copy.qxp 12/12/2017 10:59 AM Page 1 External Appeal, Internal Dominance: How Party Leaders Contribute to Successful Party Building Brandon Van Dyck ABSTRACT Many successful political parties depend for their initial popularity and cohesion, and even for their long-term brand strength, on a leader. Nevertheless, literature on successful party building downplays the role of leaders. Thus, the question, what type of leader is good for party building?, remains undertheorized. This arti - cle presents and provides initial evidence for a leadership-centered theory of suc - cessful party building. It argues that externally appealing, internally dominant lead - ers facilitate party building by lifting new parties to electoral prominence and helping to prevent debilitating schisms. The article provides evidence for this argu - ment through a most similar cases comparison of three new left parties in Latin America: two that took root (Brazil’s Workers’ Party, Mexico’s Party of the Dem - ocratic Revolution), and one that collapsed (Peru’s United Left). Keywords : Political parties, leadership, theory construction, Latin America ver the last century in Latin America, hundreds of parties have formed, but O only a tiny fraction have succeeded, or become institutionalized as major national contenders. Of this tiny fraction, many depended for their early electoral success and cohesion on a leader. In extreme cases, leaders provided the basis for enduring brands (e.g., Peronism, Chavismo ), but even in more institutionalized par - ties (Peru’s APRA and AP; Costa Rica’s PLN; Venezuela’s AD and COPEI; El Sal - vador’s ARENA, Brazil’s PT and PSDB, Mexico’s PRD), leaders proved critical for early success and survival. 1 Yet scholars generally shy away from leadership-centered explanations of party- building outcomes for fear of excessive voluntarism. Moreover, most who do focus on leaders argue that dominant figures impede party institutionalization. According to these scholars, politicians who mobilize support via personalistic appeals seldom invest in party organizations that could constrain them, and because such appeals tend to be nonprogrammatic, personalistic leaders often hinder the development of partisan brands (Panebianco 1988, 67, 147; Mainwaring and Scully 1995; Weyland 1999). This article does not dispute such claims; history offers numerous examples of leaders who have abandoned, destroyed, or seriously weakened their own parties. It does highlight, though, that dominant leaders often contribute to party building as Brandon Van Dyck is an assistant professor in the Department of Government and Law at Lafayette College. [email protected] © 2018 University of Miami DOI 10.1017/lap.2017.3 LAPS_Spring2018_LAPS_Fall13_copy.qxp 12/12/2017 10:59 AM Page 2 2 LATIN AMERICAN POLITICS AND SOCIETY 60: 1 well, and that to date, scholars have paid relatively little attention to this fact. Thus, the important question, what type of leader is good for party building?, remains undertheorized. 2 This article presents an original, leadership-centered theory of successful party building. It argues that the presence of an externally appealing, internally dominant party leader is frequently critical for new party success. Externally appealing, inter - nally dominant leaders perform two potentially vital tasks for new parties. First, they win votes. In doing so, they lift otherwise marginal new parties to electoral promi - nence. Second, they help prevent schisms: their coattails discourage defection; their preeminence prevents crippling power struggles; they facilitate decisionmaking and conflict settlement; and they rarely have incentives to defect themselves. By making new parties electorally viable and preventing schisms, externally appealing, inter - nally dominant leaders often contribute decisively to new parties’ rise and survival. Much of the article’s theory concerns where internal dominance comes from. Although electoral indispensability makes new party leaders internally powerful, it does not always make them internally powerful enough to qualify as dominant. Important factions may not prioritize vote maximization, and the leader, despite having external appeal, may lack additional assets—in particular, crossfactional ties, moral authority, and the optimal ideological profile—that are critical for maximiz - ing internal support and leverage. Consequently, while some externally appealing leaders are internally dominant, others are not. This variation can make the differ - ence between schism and survival. The article supports this argument through a “most similar cases” comparison of Brazil’s Workers’ Party (PT), Mexico’s Party of the Democratic Revolution (PRD), and Peru’s United Left (IU). The PT, PRD, and IU were factionalized, mass-based, left par - ties born during Latin America’s third wave of democratization. They depended on leaders for initial popularity: the PT on Luiz Inácio Lula da Silva, the PRD on Cuauhtémoc Cárdenas, IU on Alfonso Barrantes. Yet Lula and Cárdenas were inter - nally dominant whereas Barrantes was not, and this difference shaped the parties’ con - trasting fates—survival for the PT and PRD, collapse for the IU. Lula and Cárdenas, both popular, internally dominant leaders, lifted their parties to prominence and helped to keep them intact. By contrast, the presence of Barrantes, a popular but internally weak leader, made IU electable but vulnerable to fatal schism. Drawing on archival materials, interviews, and underutilized sources in Portuguese and Spanish, the case studies analyze why Lula and Cárdenas were more internally powerful than Barrantes and trace how this difference led to the parties’ divergent outcomes. 3 This article is an exercise in theory building. It presents a theoretical argument and provides initial evidence of the argument’s plausibility. This evidence comes in three forms: individual case studies (including a shadow case study), which illustrate causal mechanisms; explicit comparisons of the three cases; and responses to coun - terarguments. The article’s theoretical argument is not specific to left or Latin Amer - ican parties. Although the evidence in the article concerns the Latin American new left, the theoretical argument should apply to new parties across regions, periods, and the left-right spectrum. LAPS_Spring2018_LAPS_Fall13_copy.qxp 12/12/2017 10:59 AM Page 3 VAN DYCK: PARTY BUILDING 3 The article is organized in three sections. The first fleshes out the theory. The second operationalizes the variables and scores the PT, PRD, and IU. The third sets up the most similar cases comparison, presents the case studies, infers comparative lessons, addresses counterarguments, and provides initial evidence of generalizability through a shadow case study of Venezuela’s Radical Cause (LCR). THE ARGUMENT This study defines successful party building as the process by which new parties develop into electorally significant, enduring political actors. Stated differently, it is the process by which new parties become institutionalized. To be considered a case of successful party building, a new party must persist over time and consistently receive a large share of the national vote. Unsuccessful cases thus include new parties that never take off electorally and disband; those that perform well in one or two elec - tions but collapse; and those that persist over time but with a tiny share of the vote. Of the few new parties that succeed, a large number depend for early popularity and cohesion on a leader. Why? Unlike institutionalized parties, which tend to have strong partisan brands, new parties tend to have weak ones. Institutionalized parties can count on loyal voters to turn out and support them in elections, and elites in these parties are unlikely to defect because doing so would entail losing the partisan vote. But brand development is gradual; new parties must carve out distinctive iden - tities and demonstrate consistency over time (Lupu 2016). During the initial years of parties’ existence, brands are usually weak, fragile “works in progress.” Thus, new parties tend to depend on a popular leader’s coattails to garner votes and to prevent lower elites from defecting. In effect, popular leaders substitute for strong brands in many new parties. In presidential democracies especially, a new party leader’s popularity can be electorally crucial. Presidential systems compel parties to nominate politicians with broad appeal because parties without viable presidential candidates rarely become electorally competitive, and noncompetitive parties rarely endure (Samuels and Shugart 2010). In uniformly presidentialist Latin America, leaders have played a decisive role in making many new parties electorally viable. As already noted, they have laid the foundation for enduring partisan brands in some cases (e.g., Peronism, Fujimorismo , Chavismo ) and provided essential coattails in many more (Peru’s APRA and AP; Costa Rica’s PLN; Venezuela’s AD and COPEI; the Dominican Republic’s PRD and PLD; El Salvador’s ARENA, Brazil’s PT and PSDB, Mexico’s PRD). Popular leaders hardly ensure party institutionalization; indeed, they often hinder it. But without a popular leader, new parties often cannot take off (especially in presidential systems), making survival unlikely. Although popular leaders discourage defection, they provide no guarantee against schisms. In fact, new parties that electorally depend on a single leader are vulnerable to fatal schisms; after all, if the leader